ERIC EJ750800: A Descriptive Assessment of Instruction-Based Interactions in the Preschool Classroom

2006·0.14 MB·English

by  ERIC

JOURNALOFAPPLIEDBEHAVIORANALYSIS 2006, 39, 79–90 NUMBER1 (SPRING2006)
A DESCRIPTIVE ASSESSMENT OF INSTRUCTION-BASED
INTERACTIONS IN THE PRESCHOOL CLASSROOM
VIRGINIA W. NDORO
BEHAVIORANDSOCIALINTERVENTIONCENTER
AND
GREGORY P. HANLEY, JEFFREY H. TIGER, AND NICOLE A. HEAL
UNIVERSITYOFKANSAS
The current study describes preschool teacher–child interactions during several commonly
scheduledclassroomactivitiesinwhichteachersdeliverinstructions.Anobservationsystemwas
developedthatincorporatedmeasurementofevidence-basedcompliancestrategiesandincluded
the types of instructions delivered (e.g., integral or deficient directives, embedded directives,
‘‘do’’ or ‘‘don’t’’ commands), the children’s behavior with respect to the instructions (e.g.,
compliance,noncompliance,activeavoidance,problembehavior),andthedifferentialresponses
oftheteachertothechild’sbehaviorfollowinganinstruction(e.g.,appropriateorinappropriate
provisionofattentionandescape).After4classroomteacherswereobservedatleastfivetimesin
each of five target activities, simple and conditional probabilities were calculated. Results
indicatedthat(a)thefrequencyofinstructionandprobabilityofcompliancevariedasafunction
of activity type, (b) ‘‘do’’ commands and directive prompts were delivered almost to the
exclusionof‘‘don’t’’commandsandnondirectiveprompts,(c)thelikelihoodofcompliancewas
highestfollowinganembeddedoranintegraldirectiveprompt,and(d)althoughputativesocial
reinforcers were more likely to follow noncompliance than compliance and were highly likely
following problem behavior, compliance occurred over twice as much as noncompliance, and
problem behavior during instructions was very low. Implications for using descriptive
assessments for understanding and improving teacher–child interactions in the preschool
classroomare discussed.
DESCRIPTORS: descriptiveassessment,instructions,compliance,preschoolchildren,teachers
_______________________________________________________________________________
Research has shown that descriptive assess- then be developed to minimize problem
ment can be an effective means for quantita- behaviors, strengthen desirable behaviors, and
tivelydescribingimportantinteractionsbetween improve the overall effectiveness of adult–child
people, which in turn, can be used to infer interactions (Iwata, Pace, Dorsey, et al., 1994;
variables that influence behavior (Atwater & Lalli,Browder,Mace,&Brown,1993).Tothat
Morris, 1988; Bijou, Peterson, & Ault, 1968; end, several recent descriptive studies sought to
Lerman & Iwata, 1993; Thompson & Iwata, identify antecedent and consequent events
2001;Vollmer,Borrero,Wright,VanCamp,& associated with severe problem behaviors such
Lalli, 2001). Once relations between an indi- as self-injury, aggression, and disruption that
vidual’sbehaviorandthesocialenvironmentare are often exhibited by children and adults with
specified, analyses can be designed to directly developmentaldisabilities(Thompson&Iwata;
test putative relations, and interventions can Vollmer et al.).
Although descriptive assessments are often
conducted to infer variables for problem
We thank Karen Benzel for her assistance with data
behavior or to individualize subsequent func-
analysis.
RequestsforreprintsshouldbeaddressedtoGregoryP. tional analyses of problem behavior, descriptive
Hanley, Applied Behavioral Science Department, 1000 assessments also may be conducted simply to
Sunnyside Avenue, University of Kansas, Lawrence,
better understand how antecedent and conse-
Kansas66045(e-mail: [email protected]).
doi:10.1901/jaba.2006.146-04 quences interact with important behaviors in
79
80 VIRGINIA W. NDORO et al.
the natural environment. To this end, several relation to child compliance and problem
descriptive assessments have been conducted in behavior. However, the variables that we were
classrooms to identify important relations interested in measuring were primarily influ-
between teacher and child behavior (Atwater enced by developments in effective prompting
& Morris, 1988; Fagot, 1973; Lytton & and differential reinforcement strategies derived
Zwirner, 1975; McKerchar & Thompson, from the empirical literature on the treatment
2004;Strain,Lambert,Kerr,Stagg,&Lenkner, of noncompliance and escape-maintained prob-
1983). For example, McKerchar and Thomp- lembehavior.Fromthisliterature,weidentified
son described the prevalence of social con- the following strategies that appeared to be
sequences following preschoolers’ problem be- relevant to the treatment of compliance in
havior. They found that attention was provided classrooms: integral directive prompting (often
following problem behavior at least some of the referred to as three-step prompting; Horner &
time for all of the children who displayed such Keilitz,1975;Iwata,Pace,Kalsher,Cowdery,&
behavior. More relevant to instructional situa- Cataldo, 1990), embedding prompts in pre-
tions, McKerchar and Thompson found that ferred activities (Carr, Newsom, & Binkoff,
escape from instructional tasks was provided 1980; Piazza, Contrucci, Hanley, & Fisher,
following problem behavior to 33% of children 1997), framing instructions as ‘‘do’’ commands
who participated in the study, yet there did not as opposed to ‘‘don’t’’ commands (Adelinis &
appear to be a contingency between problem Hagopian, 1999; Fisher, Adelinis, Thompson,
behavior and escape in the context of an Worsdell,&Zarcone,1998;Neef,Shafer,Egel,
instruction for these children. That is, escape Cataldo, & Parrish, 1983), eliminating escape
was slightly less likely to occur following from instructions by continuing prompting
problem behavior than at other times. until the task is completed (also known as
Atwater and Morris (1988) also conducted escape extinction; Iwata, Pace, Cowdery, &
descriptive assessments in preschool classrooms, Miltenberger,1994;Iwataetal.,1990;Zarcone
but their analysis was restricted to instructional et al., 1993), and providing social positive and
situations. These authors measured teacher negative reinforcers for compliance (Lalli et al.,
behavior in the form of an instruction (sugges- 1999; Parrish, Cataldo, Kolko, Neef, & Egel,
tion, imperative, question, or declarative) and 1986; Piazza, Fisher, et al., 1997).
the type of verbal feedback (approval or Although the effects of these strategies have
disapproval), the context in which an instruc- been experimentally demonstrated with chil-
tion was delivered (small or large group, dren and adults who display various topogra-
transitions), and children’s behavior (compli- phies of problem behavior and noncompliance,
ance, off task, and disruptions). Their results it is unclear whethertherelations betweenchild
suggested that the form of the instruction did behavior (compliance and problem behavior)
not influence the probability of compliance as and the experimentally demonstrated strategies
much as the interaction context in which the would hold under nonexperimental conditions,
instruction was delivered (e.g., children were and whether these strategies could be imple-
more likely to comply with an instruction if mented by early childhood teachers, especially
they were engaged in an activity than if they those working with typically and atypically
were off task or disruptive). developing children in the preschool classroom.
Theaimsofthepresentstudyweresomewhat Their adoption seems important as a preventive
similar to those of Atwater and Morris (1988) strategy aimed at preempting the development
in that we also sought to measure and describe of early patterns of noncompliance and escape-
aspects of teachers’ instructions and their related aberrant behavior. Therefore, this study
INSTRUCTION-BASED INTERACTIONS 81
offers a method of descriptive assessment of instructions); however, 6 of the children had
instruction-based interactions in the preschool limited expressive language skills (e.g., they
classroom. More specifically, we measured communicated with gestures and single-word
aspects of teacher-mediated instructions and utterances).
consequences—those that have repeatedly been Observations were conducted in three loca-
showntoalterthelikelihoodofcomplianceand tions—in the main classroom (36 m by 20 m),
escape-maintained problem behavior with indi- on theplayground (85 m by110 m),and in an
viduals who display noncompliance and severe indoor activity room (23 m by 30 m)—during
problem behavior—in a university-affiliated five typically scheduled activities (free choice,
preschool classroom. circle time, meals, outdoor, and centers). Free
choice was a 45-min period in which children
independently selected from one of nine
METHOD
simultaneously available activities (dramatic
Participants and Setting
play, blocks, art, games, computer, etc.). Circle
Naturalisticobservationswereconductedtwo time was a 15-min teacher-led activity during
to four times a day in an inclusive, full-day whichchildrensatinahalf-circlefacingthelead
preschool classroom serving 15 children. teacher. The lead teacher engaged the children
Trained graduate students intermittently ob- in songs, finger plays, conversations, or had
served 4 undergraduate teachers who were them participate in a science, craft, or cooking
enrolled in an early childhood teacher-training activity. Family-style dining was arranged
program. These teachers attended a half-day during meals. Small groups of children sat with
orientation session prior to interacting with the a teacher at a small table, children passed and
children in the classroom; this session included served food and beverages, and the teacher
a lecture on the importance of integral directive modeled appropriate mealtime behavior (e.g.,
prompting, embedding instructions, ‘‘do’’ ver- washing hands prior to eating, appropriate use
sus ‘‘don’t’’ instructions, escape extinction, and of silverware, chewing with mouth closed).
differential reinforcement of compliance. A Outdoor time was a 45-min period during
trained graduate student provided feedback on which various physical activities including
the correct implementation of these skills balancing, tumbling, running, skipping, climb-
during a 2-week period prior to the onset of ing, and bike riding were encouraged. During
data collection while the teachers interacted centers, children rotated every 8 min between
with the children in the classroom and different teacher-structured activities that in-
following their respective shifts. At least 2 of cluded manipulative/table-top skills (emphasis
the 14 teachers were present during each wasonfinemotorskillssuchasblockbuilding),
descriptive observation session. art skills (exploration and creativity activities
Fifteen children between 30 and 48 months suchascutting,drawing,pasting,clayforming),
of age participated. Twelve of the children were concept skills (matching, pointing to, or
typically developing, and English was a second naming colors, shapes, alphabet, identifying
language for 3 of them. Two children had been body parts), and small-movement skills (cut-
diagnosed with nonspecified developmental ting, tracing, and writing were taught and
delay and were receiving intermittent support practiced).
services from the local school district, and 1
child had been diagnosed with autism and was Response Measurement and
receiving one-on-one support throughout the Interobserver Agreement
school day. All children had some receptive Observations were conducted during 15-min
language skills (e.g., they could follow simple periods. A data sheet partitioned into 1-min
82 VIRGINIA W. NDORO et al.
intervals was used to record the occurrence of directive was made. An integral directive was
teacher and child interactions. A discontinuous scored if a fixed hierarchy of vocal, model, and
observation procedure was used in which the physical prompts were issued, with 3 to 5 s
first instructional episode initiated by the target between prompts to allow compliance, and if
teacher in each 1-min interval was recorded. theinitial(vocal)promptspecifiedtheactionto
The observer circled specified behavior codes in be completed or terminated. A deficient di-
three categories of events arranged sequentially. rective was scored if the teacher did not include
In essence, a close-ended antecedent-behavior- a fixed hierarchy of vocal, model, and physical
consequence observation system was used to prompts (e.g., subsequent prompts were not
describe instruction-based interactions. Teacher deliveredfollowing3 sofnoncompliance),ifan
instructionswerecodedfirstandincludedcodes action (or termination of an action) was not
for ‘‘do’’ and ‘‘don’t’’ commands, integral specified, or if the prompt was phrased as
directives, deficient directives, and embedded a question (e.g., ‘‘Can you come here?’’ or
directives. The second category was child ‘‘Would you sit down?’’). An embedded di-
behavior and included compliance to the vocal rectivewasscoredifaninstructionwasdelivered
prompt, compliance to the model prompt, withoutadirectstatementofaspecificgoal,was
noncompliance, active avoidance, and problem embedded in a play activity, and, if completed,
behavior. The consequent event category was accomplished a teacher’s goal (e.g., ‘‘Let’s hop
located third, consisted of the teacher’s re- like bunnies to the bathroom’’). These three
sponses to child behavior, and included appro- categories of instructions (integral, deficient, or
priate attention, inappropriate attention, ap- embedded) were also mutually exclusive cate-
propriate escape, and inappropriate escape. gories (i.e., one, and only one, was scored for
Each teacher was individually observed at least each instructional trial).
fivetimesineachofthefiveactivities.Atotalof Children’s behaviors were scored following
114 15-min observations were conducted for the onset of an instruction, and were opera-
a total of 28.5 hr of teacher–child observation. tionally defined as follows. Compliance vocal
The first instruction delivered by the target was scored on the completion of an instructed
teacherineach1-minintervalinitiatedthedata- response within 5 s of the teacher’s initial vocal
collection sequence. Data collection continued prompt, and compliance model was scored on
with regards to that particular instructional completion of aninstructed response within5 s
sequence until either the instruction was of the teacher’s model prompt. Noncompliance
completed or the teacher allowed the child to was scored if the instructed response was not
escape the instruction. This usually occurred completed within 5 s of the teacher’s second
within the same interval, but not always. In the prompt.Activeavoidancewasscoredifthechild
latter case, observers were instructed to record ran away, crawled under the table, or fell to the
a new instructional sequence at the start of the floor following an instruction from the teacher.
subsequent interval. We first recorded whether The problem behavior category was scored if
the initial instruction specified some action the child exhibited any of the following
(e.g., ‘‘Hand me the book’’), which were scored behaviors after an instruction had been de-
as ‘‘do’’ commands, or whether the instruction livered: self-injurious behavior (banging or
specified a behavior to be terminated or not hitting head, biting self), aggression (hitting,
emitted (e.g., ‘‘Don’t throw the book’’), which kicking scratching, pinching, biting others),
were scored as ‘‘don’t’’ commands. These were disruption (knocking down materials, pushing
mutuallyexclusivecategories. Astheinstruction away chairs or tables), and inappropriate
unfolded, a determination of the type of vocalizations (screaming, swearing). The com-
INSTRUCTION-BASED INTERACTIONS 83
Table 1
pliancemeasures(compliancevocal,compliance
InterobserverAgreementPercentagesacrossTeachers’and
model, and noncompliance) were mutually
Children’s Behavior
exclusive categories. By contrast, the active
avoidance and problem behavior could be Behavior M Range
scored at any time during the instructional
Teachers’instructions
episode and in addition to our compliance ‘‘Do’’instructions 93 80–100
‘‘Don’t’’instructions 100 100
measures.
Integraldirectives 88 60–100
The final set of behaviors recorded were the Deficientdirectives 87 60–100
Embeddeddirectives 99 93–100
teachers’ responses to child behavior (which
Children’sresponses
followed the teacher’s instruction). Either Compliancetotheverbalprompt 87 67–100
Compliancetothemodelprompt 98 80–100
appropriateattentionorinappropriateattention
Noncompliance 88 67–100
was recorded for each instructional episode. Activeavoidance 99 93–100
Problembehavior 100 100
Appropriateattention was scored if any form of
Consequencesprovidedbyteacher
attention was provided following compliance or Appropriateattention 84 60–100
ifallformsofattentionwerewithheldfollowing Inappropriateattention 80 60–100
Appropriateescape 100 100
noncompliance, active avoidance, and problem Inappropriateescape 92 80–100
behavior(additionalprompting,i.e.,amodelor
physical prompt, was not scored as attention).
was defined as both observers recording the
Inappropriate attention was scored if any form
occurrence of the same events within the same
of attention was provided following noncom-
1-min interval. These scores were summed
pliance, active avoidance, or problem behavior
across intervals, divided by the total number
or if all forms of attention were withheld
of agreements plus disagreements, and multi-
following compliance. Appropriate escape was
plied by 100%. Mean interobserver agreement
scored if instructions were terminated or
across behaviors is reported in Table 1.
continued following compliance or were with-
held following noncompliance, whereas inap-
RESULTS
propriate escape was scored only if an in-
struction was terminated following non- The descriptive assessment involved 114 15-
compliance. The category of appropriate escape min observations (28.5 hr of observation).
included more teacher responses than inappro- Instructional instances were captured in 947
priate escape because it was determined that (55%) of the 1,710 1-min time samples. The
both terminating or continuing instructions meannumberandrangeofinstructionsandthe
following compliance were appropriate teacher overall percentage of compliance with instruc-
responses, considering the routine nature of tions in each of the five activity areas are
both singular instructions and instructional reported in Table 2. There was a small
sequences often provided in the classroom. difference in the mean number of instructions
Several graduate students in a child psychol- delivered across activities with the teacher-led
ogy program were trained on the observational activities (centers and circle) associated with
code until they each attained 80% agreement a greater number of instructions (9.3 and 8.8,
scores for each of the 14 behaviors (teacher and respectively) than the more child-directed
child) for three consecutive sessions. Interob- activities (outdoor and free choice; 8 and 7.2,
server agreement was then assessed by having respectively). The percentage of compliance
two observers collect data simultaneously but with instructions was similar for centers
independently during 20% of the sessions (68%), meal (66%), and free choice (65%),
across teachers and activities. An agreement but higher levels of compliance were observed
84 VIRGINIA W. NDORO et al.
Table2
the inappropriate provision of escape (i.e.,
Mean Number (and Range) of Instructions and the
escape provided following problem behavior,
Overall Percentage of Compliance to Instructions (per
active avoidance, or noncompliance) was ob-
15-min Observation) during Various Preschool Activities
served on fewer occasions (20%) than the
Meanandrangeforthenumber Percentage appropriate management of escape (i.e., escape
Activity ofinstructionsdelivered ofcompliance
withheld following problem behavior, active
Centers 9(4–15) 68 avoidance, or noncompliance or provided
Circle 9(4–14) 84
following compliance, 80%).
Meal 8(3–11) 66
Outdoor 8(4–11) 59 Figure1showsthesimpleprobabilitiesofthe
Freechoice 7(1–13) 65
various teacher prompts, child behaviors, and
teacher consequences during the instructional
during circle (84%) and lower levels of
episodes, and Table 3 shows the conditional
compliance were observed during the outdoor
probabilities of various child behaviors (e.g.,
activity (59%).
compliance, problem behavior) given the vari-
The overall number of instances (in paren-
ous types of teacher instructions. Similarly,
theses)andsimpleprobabilities(verticalbars)of
Table 4 shows the conditional probability of
teachers’ instruction types, child behaviors, and
various teacher-delivered consequences (e.g.,
teacherresponsesgivenaninstructionalinstance
attention, escape) given the various types of
are depicted in Figure 1. The top panel shows
child behaviors (e.g., noncompliance, active
that ‘‘do’’ instructions were observed during
avoidance). The data in Table 4 were extracted
99% of all instructional instances, whereas
by comparing the specific child behaviors and
‘‘don’t’’instructionswererarelyobserved(seven
the types of consequences that were scored. For
instances). The majority of directive prompts
instance,iftheonlychildbehaviorrecordedwas
(60%) were deficient (i.e., insufficient or
noncomplianceandinappropriateattentionwas
extended delay between prompts, lack of
recorded, then we counted that trial as one in
follow-through with a model or physical
which attention was provided for noncompli-
prompt), and only five directives were embed-
ance with an instruction.
ded into a play-based activity.
As can be seen in Table 3, ‘‘do’’ directives
Althoughproblembehaviorandactiveavoid-
resulted in somewhat higher levels of compli-
ance were rarely observed during instructional
episodes in this classroom (total of 15 episodes; ance than ‘‘don’t’’ directives (68% and 57%,
see middle panel of Figure 1), noncompliance respectively). However, because ‘‘don’t’’ direc-
to instructions was observed on 301 occasions tives were observed only seven times, the
(i.e., 32% of instructions did not result in stability of the percentage of compliance
compliance).Whencompliancewasobserved,it associated with ‘‘don’t’’ directives remains in
was usually following the initial vocal prompt. question. Similarly, embedded directives were
The bottom panel of Figure 1 shows that delivered only five times, so it remains unclear
attention was provided following problem whether the high levels of compliance (100%)
behavior, active avoidance, or noncompliance associated with this type of instruction would
or was withheld following compliance during continue if more embedded directives were
the majority of instructional episodes (‘‘in- delivered.
appropriate attention,’’ 64%). Conversely, at- By far, the majority of instructions delivered
tention was provided following compliance and wereeitherintegraldirectiveprompts(n5373)
was withheld following problem behavior, ordeficientdirectiveprompts(n5569),sothe
active avoidance, or noncompliance during percentages of compliance associated with these
36% of the instructional episodes. By contrast, two types of instruction are probably quite
INSTRUCTION-BASED INTERACTIONS 85
Figure 1. The percentage and number of instruction types (top), child behaviors (middle), and teacher-delivered
consequences (bottom) observedacross allinstructionalinstances.
86 VIRGINIA W. NDORO et al.
Table 3
The Conditional Probabilities ofChildren’s Responses toDifferentTypes of Teacher Instructions (the Numbersin
Parentheses Depict the Overall Numberof Instances Recorded)
Givenadirective Givenadirective Givenanintegral Givenadeficient Givenanembedded
Conditionalprobabilityof: ‘‘do’’(940) ‘‘don’t’’(7) directive(373) directive(569) directive(5)
Compliance(646) 68%(642) 57%(4) 87%(323) 56%(318) 100%(5)
Noncompliance(301) 32%(298) 43%(3) 13%(50) 44%(251) 0%(0)
Problembehaviororactive 1%(12) 43%(3) 1%(3) 2%(12) 0%(0)
avoidance(15)
stable. Compliance was much higher when as some form of attention (other than that
integraldirectivepromptswere delivered(87%) provided by additional appropriate prompting)
than when deficient directives were given was provided following 61% of the instances of
(56%). noncompliance (i.e., inappropriate attention).
Although problem behavior was quite low at In addition, escape was provided more often
15 episodes, at least one interesting association following noncompliance (65%) than it was
was evident (see the bottom row of Table 3). withheld (35%). In essence, putative social
That is, problem behavior was considerably positive and negative reinforcement in the form
more probable following a ‘‘don’t’’ directive of attention and escape favored noncompliance
(43% of ‘‘don’t’’ directives were followed by over compliance. A similar relation is evident
problem behavior) relative to the other types of for problem behavior, in that attention was
instructions (less than 2% of each of the other much more likely to be provided following
types of instructions were followed by problem problem behavior (87%) than withheld follow-
behavior). The probability of this high percent- ing problem behavior (13%). By contrast, there
agefor‘‘don’t’’directivesoccurringbychanceis was an equal probability that problem behavior
quite low (Z 5 8.7; p , .001). It also may be would result in the provision or prevention of
worth noting that 12 ‘‘do’’ directives were escape.
associated with problem behavior, and of these,
nine (80%) were deficient directives and three
DISCUSSION
(20%) were integral directives. However, this
differencewasnotstatisticallysignificant(dueat The methods of descriptive observation out-
least in part to the low levels of problem lined above allowed us to identify the naturally
behavior and the relatively high levels of occurring probability of both antecedent and
deficient prompts). consequent events that have been experimen-
As can be seen in Table 4, attention was tally verified as having an impact on children’s
providedfollowingonly35%oftheinstancesof behavior during instruction. Although the
compliance (i.e., appropriate attention), where- generality of these findings is limited by the
Table 4
The Conditional Probabilities ofTeacher-Delivered Social Consequences for Children’s Behavior
Givencompliancewith Givennoncompliancewith Givenproblembehavioror
Conditionalprobabilityof: aninstruction aninstruction activeavoidance
Attention 35% 61% 87%
Noattention 65% 39% 13%
Escape 65% 50%
Noescape 35% 50%
INSTRUCTION-BASED INTERACTIONS 87
uniqueness of the setting (university-based pre- Instead of high levels of teacher instruction and
school), teachers(young adults seekingbachelors direction, free-choice periods are typically de-
degrees in early childhood education), and their scribed as teachers initially setting the occasion
training(eachhadsomebehavior-analyticcourse for learning by attractively displaying materials
work and attended an orientation about the and then commenting on children’s use of
importance of effective prompting and differen- language and play with respect to the materials.
tial reinforcement), the descriptive observations The current results suggest that child-initiat-
did yield some interesting patterns. edactivitiesmayinvolvemoredirectprompting
Theinitialquestionsaddressedintheanalysis from teachers than previously thought (at least
were whether different activities were associated that was the case in this setting). It is also
with more or fewer instructions or greater or possible that the current participants (i.e.,
lesser probabilities of compliance. Although the student teachers) had not received sufficient
mean number of instructions was higher in training or practice in organizing and conduct-
teacher-led activities (centers and circle) than in ing child-initiated activities. Future research
child-initiated activities (outdoor and free should be directed toward determining whether
choice), the difference between the amount of the current findings were specific to these 4
instructions delivered across teacher- and child- studentteachersorwhetherteachers,ingeneral,
initiated activities was small (an average of 2.1 providetoomanypromptsduringactivitiesthat
instructions per 15-min period). This finding are supposed to promote child initiations and
was somewhat surprising given prior descrip- choices. At a minimum, these results suggest
tions in the literature regarding differences that the current teachers should receive direct
betweenteacher-ledandchild-initiatedactivities teachertrainingininteractionsrelevanttochild-
(Delprato, 2001; Losardo & Bricker, 1994; initiated activities (e.g., commenting or engag-
Wolery & Sainato, 1996). ing a child in joint attention to an activity,
Teacher-led activities often involve assess- rather than providing direct instructions).
ment and instruction with respect to a wide Additional research is needed to determine
variety of curriculum-based developmentally whether most teachers require more direct
appropriate skills (e.g., matching, pointing, training in organizing and conducting child-
and naming colors, numbers, shapes), which initiated activities than is currently provided in
incorporatematerials thatareof generalinterest typical teacher-training programs.
to young children (Essa, 2003). Social skills It was notable that circle was associated with
such as waiting for one’s turn, listening to the highest percentage of child compliance
others, and attending to a speaker may also be (84%) and outdoor time was associated with
assessed and taught in these group activities. By the least (58%), whereas the level of compliance
contrast, child-initiated activities (free choice, in the other three areas (centers, meals, and free
free play, guided discovery) involve the simul- choice) did not differ much (all percentages in
taneous availability of several activities and the mid 60s).Atwater and Morris(1988) found
materials from which the child can choose. that compliance was more likely when an
Children have opportunities to develop and instruction was delivered to a child engaged in
practice skills such as painting, drawing, ataskratherthanifthechildwasbeingdisruptive
manipulating toys, block building, and playing or off task. The present results also suggest that
gamesattheirownpace.Socialskillsrelevantto the activity context may be associated with
peer relations also may be developed during varying probabilities of compliance.
these times (Allen & Schwartz, 2001; Brede- Although the results of the present study and
kamp & Copple, 1997; Goetz & Allen, 1983). those of Atwater and Morris (1988) were
88 VIRGINIA W. NDORO et al.
consistent with respect to the influence of problem behavior), and escape was more
context, the results of the two studies differed probable for noncompliance (relative to com-
regarding the influence of instruction form. pliance), yet levels of compliance were far
Atwater and Morris found that the form of the higher than levels of noncompliance and
instruction was not related to the probability of problem behavior in the classroom.
child compliance. By contrast, in the current The reason for these somewhat counterintu-
study, embedded and integral directives were itive findings is not entirely clear, but there are
associated with high levels of compliance several possible explanations that could be the
(100% and 87%, respectively), and ‘‘don’t’’ focus of future research. First, it is possible that
promptsanddeficientdirectiveswereassociated the form of the instructions exerted antecedent
withlowerlevelsofcompliance(57%and56%, control sufficient to yield higher levels of
respectively). The two studies may have pro- compliance despite contingencies that generally
duced discrepant results because different in- supportednoncompliance.Second,itispossible
structional forms were measured in each. that the consequences delivered by the teachers
Atwater and Morris distinguished forms based had effects different from those suggested by
on their syntactic structure (e.g., direct im- previous research (Lalli et al., 1993, 1999;
peratives, ‘‘let’s’’ imperatives, questions, and Parrishetal.,1986;Piazza,Fisher,etal.,1997).
declaratives), whereas the different forms mea- For example, it is possible that different forms
sured in the present study were based on ofattentionweredeliveredforcompliance(e.g.,
existingempiricalevidenceofarelationbetween praise) than for problem behavior (e.g., repri-
compliance probabilities and particular instruc- mands),andthatthesequalitativeparametersof
tion forms (e.g., ‘‘do’’ vs. ‘‘don’t,’’ integral vs. reinforcement may have overridden the relative
deficient). rates of reinforcement delivered for compliance
Several studies have shown that compliance and problem behavior (e.g., Neef & Lutz,
andproblembehavioroftencovaryinaninverse 2001). A related possibility is that the activities
relation, and that targeting one can produce were usually preferredbythechildren,in which
concomitant but opposite changes in the other case termination of the activity contingent on
(Iwata et al., 1990; Lalli et al., 1999; Parrish et problem behavior may have functioned as
al., 1986; Piazza, Fisher, et al., 1997). In light punishment (i.e., time-out from positive re-
ofthesepreviousstudies,itisnotsurprisingthat inforcement) rather than escape from nonpre-
levels of compliance were relatively high (68%) ferred activities. Nevertheless, further increases
and levels of problem behavior were quite low in compliance (from compliance with two of
(2%).However,itissurprisingthatthelevelsof every three instructions observed in the current
compliance were so much higher than problem study) may be gained by training teachers to
behavior, given that the teacher-delivered con- more carefully allocate potential social reinforc-
sequences seemed to favor the opposite (higher ers exclusively for desirable behaviors (e.g.,
levels of problem behavior and lower levels of compliance, on task, etc.) in addition to
compliance; see the bottom panel of Figure 1 delivering instructions in an optimal format
and Table 4). Inappropriate attention (i.e., (e.g., issuing ‘‘do’’ as opposed to ‘‘don’t’’
delivery of attention for noncompliance or instructions, and delivering integral directives).
problem behavior and the nondelivery of Several limitations of the current investiga-
attention following compliance) was more tion are worth noting. First, the data are
prominent than appropriate attention (i.e., aggregated across several preschool teachers as
delivery of attention for compliance and the they interacted with multiple children. There-
nondeliveryofattentionfornoncomplianceand fore, therelations thatare impliedbythese data